Projections of Antarctica’s contribution to future sea-level rise span a wide range. This talk addresses a narrower question: to what extent does the present-day behaviour of the Antarctic Ice Sheet contain information about its future dynamic mass loss? Using ensembles of ice-sheet simulations, we quantify the relationship between contemporary or near-term mass loss and subsequent contributions across drainage basins and prediction horizons.

We find that predictability is strongly regional and time dependent. In most basins, predictive memory weakens substantially beyond a few decades, implying that agreement with present-day observations does not, by itself, tightly constrain end-of-century outcomes. The Amundsen Sea sector is a notable exception: present mass loss remains related to future contributions on centennial timescales, offering greater scope for observations to narrow projections for 2100.

These results do not imply that ice-sheet models are uninformative. Rather, they distinguish the ability to simulate physically plausible futures from the ability to constrain a particular future trajectory from the current state. This perspective supports regionally differentiated priorities for sustained observations, process understanding, high-resolution modelling and communication of projection uncertainty.